The effect of changing hematocrit (Hct) on solute removal during high efficiency hemodialysis was evaluated in 12 patients. In five subjects, Hct was raised by recombinant human erythropoietin (rHuEPO) treatment, and in the other seven by blood transfusion. Solute removal was assessed by measuring: (1) whole blood (kb), blood water (kbw) and dialysate (kd) clearances; (2) the amount of solute in the spent dialysate; (3) the fractional decrement of serum solute concentration achieved by hemodialysis; and (4) urea kinetics, including kt/V and protein catabolic rate (PCR). The results showed that increasing the Hct did result in a slight reduction in some solute clearances. The decrement, however, was minor (5 to 8%), whereas the rise in Hct was marked (55 and 65%) in the transfused and EPO-treated groups, respectively. More importantly, linear regression analysis of kd/kb ratios versus Hct indicated that a rise of Hct from 20 to 40% would reduce creatinine and phosphate clearance by 8 and 13%, respectively. By contrast, assessment of the absolute amount of solute removed in the spent dialysate failed to detect differences between the two study periods. Additionally, a rise in Hct also did not affect urea kinetic parameters including kt/V and PCR. Based on these data, it appears prudent to increase hemodialysis prescription by 10 to 15% when Hct is raised to near 40% to avoid excessive retention of molecules with slow transcellular movement.